DETROIT – Patients with proximal pancreatic cancer should be reimaged before surgery if more than 3 weeks have passed since their most recent cross-sectional imaging study.
The recommendation is based on a retrospective analysis involving 487 patients that identified a significant, roughly twofold increase in unanticipated metastasis encountered at surgery if the interval between imaging and operation was more than 3 weeks.
Dr. Joshua A. Waters
Among 293 patients with proximal pancreatic cancer and precise imaging data, the frequency of occult metastasis was 12% when the imaging-to-operation interval (IOI) was 20 days or fewer, compared with 20% at an IOI of 21-27 days, 25% at 28-34 days, 35% at 35-41 days, 29% at 42-48 days, and 30% at 49-86 days.
There were no significant differences between patients in the various 1-week intervals in terms of sex, age, tumor size, percentage of poor grade tumors, lymph node positivity, vascular invasion, or perineural invasion, Dr. Jeffrey Glant said at the annual meeting of the Central Surgical Association.
In linear regression analysis, the relationship between frequency of unanticipated metastases and weekly IOI was statistically significant (P = .006) and had a correlation coefficient R2 value of 0.99.
A similar relationship was not observed among 36 patients with distal pancreatic cancer and precise imaging data. The frequency of unanticipated metastasis among these patients was 0% at 0-6 days, 33% at 7-13 days, 38% at 14-20 days, 0% at 21-27 days, 0% at 28-34 days, 20% at 35-41 days, 33% at 42-48 days, and 25% at 49-87 days, said Dr. Glant of the department of surgery at Indiana University in Indianapolis.
Cross-sectional imaging is the primary preoperative staging modality in pancreatic cancer, which is the fourth most common cause of cancer death in the United States. The rate of encountering metastasis at operation is typically 10%-30%, he noted. The reasons for the delay in surgery could not be ascertained from the retrospective data.
Dr. Glant reported on 487 patients undergoing planned pancreatic resection for pancreatic ductal adenocarcinoma between January 2004 and December 2009 at the university’s high-volume pancreatic surgery center. Patients were excluded if they had received neoadjuvant therapy, prior pancreatic resection, or exploratory surgery for suspected metastatic disease.
Precise imaging data were available for 329 patients. Cross-sectional imaging was defined as dual-phase, contrast-enhanced CT, or MRI if CT was contraindicated.
Of the 285 patients (59%) who had their most recent imaging study performed at the university, 202 underwent resection and 83 were not resected. Metastasis was discovered at time of operation in 39 patients, he said.
Of the 202 (41%) patients whose most recent imaging study was performed at an outside institution, 139 underwent resection and 63 were not resected. Among the 202 patients, 35 had metastasis discovered at time of operation.
The overall frequency of unanticipated metastasis was statistically similar between patients who were imaged at the university and those who were imaged at an outside institution (14% vs. 17%), Dr. Glant said. This was true whether the patients had proximal (14% vs. 17%) or distal (15% vs. 25%) disease. Patients imaged at an outside venue, however, had significantly larger tumors than did those imaged at the university (3.4 cm vs. 3.1 cm; P = .05) and a higher rate of vascular invasion (73% vs. 61%; P = .03).
"It is appropriate and advisable to obtain more current imaging if the delay [in surgery] will exceed 3 weeks," Dr. Glant said.
Invited discussant Dr. Carl R. Schmidt of the Ohio State University, Columbus, said the size of the series, the robustness of the analysis, and the time period evaluated left no doubt in his mind about the use of modern imaging and the validity of the main finding. He asked what proportion of patients at the university undergoes staging laparoscopy, and what the distribution of metastasis was.
Coauthor Dr. Joshua A. Waters, also of Indiana University, replied that with just 16 patients undergoing staging laparoscopy in the entire series, the procedure is not routinely performed at the university in the setting of proximal pancreatic cancer. He also cited a recent study reporting a decrease in yield of staging laparoscopy from 1995 to 2005, particularly in those with proximal pancreatic cancers (J. Am. Coll. Surg. 2008;206:445-50). Regarding the distribution of the metastases, 80% were on the liver and 20% were at another location, primarily peritoneal implants, he said.
When asked whether any patients had undergone endobiliary stenting prior to resection, Dr. Waters said that a significant proportion of patients were stented prior to arrival at the hospital, and that stenting is known to affect the sensitivity of staging in terms of cross-sectional imaging.